Forecasting the Cost of Financing ART in Sub-Saharan Africa under Differential Funding Scenarios Tahvi Frank and Austin Carter July 25, 2018

Context HIV Financing in Sub-Saharan Africa Stagnating government health expenditure and decreasing development assistance to health Current funding levels do not meet treatment needs 2017 Purchasing Power Parity (Billions) Year Spending on health and HIV/AIDS: domestic health spending and development assistance in 188 countries, 1995–2015 Dieleman, Joseph L et al. The Lancet , Volume 391 , Issue 10132 , 1799 - 1829

Predicting Future ART Cost Predicting cost of financing ART 2017 – 2040 Estimate future ART coverage under better, worse, and reference scenario Price of ART Health spending Model the HIV epidemic for each scenario ART Price ART Coverage Population Projection Health Expenditure (better, worse, reference)

Forecasting ART Price Data on average cost of ART per person per year from Global Price Reporting Mechanism spanning 2004-2016 Stochastic frontier analysis bounds future minimum cost Gaussian process regression used to complete time series and extend to 2040 for each country Range of Annual ART Price SSA Annual ART Price per Person (US$) Year

Projected Health Expenditure Calculated annualized rate of change of per capita health expenditure in each country Development assistance for health (DAH) for HIV Government health expenditure as source (GHES) 15th (worse), 50th (reference), and 85th (better) percentile rate of change across countries used to project future funding Per Capita Spending (US$) Year

Predicted ART Coverage Malawi Scenario – Specific ART Coverage Percent Covered Year

Predicted Incidence Counterfactual trend in incidence forecasted using 15th, 50th, and 85th percentile rate of change approach Final projected incidence decreases with improvements in ART coverage along with changes in the secular trend in counterfactual incidence Predicted at country-, age-, sex-, year-, draw-level

Spectrum Rate (per 1)

HIV Prevalence Predicted HIV Prevalence Rate, 2040 Worse Reference Rate (per 1) Predicted HIV Prevalence Rate, 2040 Better Reference Worse

People Living With HIV Malawi Sub-Saharan Africa

Cost of Treatment People Living with HIV (millions) Location 2016 2040 Worse Reference Better Sub-Saharan Africa 26.1 81.2 40.1 29.1 South Africa 6.4 14.1 7.5 5.9 Nigeria 3.3 11.2 5.4 3.7 Tanzania 1.6 6.3 3.1 2.2 Cost of ART for all PLHIV, 2040

Conclusions Despite projected declines in incidence rate, the number of people living with HIV is predicted to increase Driven by demographic changes and longer lifespans for PLHIV Highlights need for increased funding Though the number of PLHIV is projected to increase across all three scenarios, worse scenario reveals dramatic consequences of reducing funding into the future Triple the number of PLHIV compared to 2016 Better scenario evidences potential for sustained increases in funding to lead to decreased prevalence and future savings on treatment Limitations Prediction of price of drug rather than comprehensive cost of care Lack of data on government health expenditure specific to HIV Analysis focuses on relationship between funding and ART coverage Does not include PrEP, male circumcision, and other highly effective interventions

Forecasting the Cost of Financing ART in Sub-Saharan Africa under Differential Funding Scenarios Tahvi Frank and Austin Carter tahvif@uw.edu aucarter@uw.edu